Quasarch Funding Proposal - Go Client Library

[cloud-j-luna]

Summary

We are proposing a funding request of 25k USD in $AKT from the community pool for our efforts in developing the Akash Network Go Client Library.

Introduction

Integration with the Akash Network is currently very challenging for third parties. A Javascript Client Library is available that powers the web integrations built on Akash Network. For services and tools built using other technologies there is no standard way to integrate with he Akash Network which is very limiting in the growth of the Akash Network developer ecosystem.

Go Client Library

The Go Client Library offers Go developers (a big percentage of cloud-native developers) a way to interact with the Akash Network. It allows for querying deployments, providers, bids, leases as well as creating, updating and closing deployments and all related tasks (check the specification for more details). The current version of the library implements all the low level operations required to fully operate the Akash Network.

Having such a library available will allow the reimplementation of critical products such as a better CLI for the Akash Network. A new version for the Terraform Provider that removes the dependency of having Akash installed. The libraries are also being currently used by our team to build new amazing software on top of Akash Network (more about this in the future).

Funding Request

We request funding of USD 25k in $AKT by next month to open source the implemented solution.

Funding Breakdown

15k - Open source the codebase

Hourly rate $34
An Estimate of 20 hours a week for ~22 weeks
A total of $14960

10k - First 6-months maintenance and support to the community and new features (not core to the spec)

Hourly rate $20
An Estimate of 20 hours a week for ~26 weeks
A total of $10400

What are these new features?

New feature include anything that the community deems necessary to have a good experience using the library. For instance having the ability to setup an Akash account or issuing certificates. These are not part of the 1.0 spec but we are open to implement them if deemed necessary during the maintenance window/budget.

Conclusion:
Implementing these solutions will significantly enhance the Akash Network's functionality and provide a better developer experience to ecosystem developers. It will allow a new set of systems to emerge in the ecosystem and unlock new types of automations.

[cloud-j-luna]

Code examples of the library usage:

Querying a deployment

func main() {
	akash, err := client.NewAkashClient()
	if err != nil {
		panic(err)
	}

	req := akttypes.QueryDeploymentRequest{
		ID: akttypes.DeploymentID{
			Owner: "<addres>>",
			DSeq:  12345,
		},
	}

	resp, err := akash.GetDeployment(context.Background(), req)
	if err != nil {
		log.Fatalf("error: %v", err)
	}

	log.Printf("%+v", resp.Deployment)
}

Querying providers

func main() {
	akash, err := client.NewAkashClient()
	if err != nil {
		panic(err)
	}

	req := v1beta3.QueryProvidersRequest{}

	resp, err := akash.GetProviders(context.Background(), req)
	if err != nil {
		log.Fatalf("error: %v", err)
	}

	log.Printf("%+v", resp.Providers)
}

Creating Deployment

func main() {
	akash, err := client.NewAkashClient()
	if err != nil {
		panic(err)
	}

	f, err := os.Open("path/to/sdl.yaml")
	if err != nil {
		panic(err)
	}

	opt := tx.CreateDeploymentOptions{
		Deposit: helpers.DefaultDeposit,
		Options: tx.Options{
			AccountName:    "<account>",
			AccountAddress: "<address>",
			Backend:        "<backend>",
		},
	}

	msg, err := helpers.BuildMsgCreateDeployment(f, opt)
	if err != nil {
		panic(err)
	}
	if _, err := akash.CreateDeployment(context.Background(), *msg, opt.Options); err != nil {
		log.Fatalf("error: %v", err)
	}
}

Querying Bids

func main() {
	akash, err := client.NewAkashClient()
	if err != nil {
		panic(err)
	}
	bidReq := v1beta3.QueryBidsRequest{Filters: v1beta3.BidFilters{
		DSeq:  12345,
		Owner: "<address>,
	}}
	bids, err := akash.GetBids(context.Background(), bidReq)
	if err != nil {
		log.Fatalf("error: %v", err)
	}
	if len(bids.Bids) <= 0 {
		log.Fatalf("No bids")
	}
}

[troian]

Hi @cloud-j-luna, thanks for making the proposal and providing examples.

What improvements it brings over existing akash-api and client. Eventually, the client from the node repo will be moved to akash-api and I'm wondering if there is no duplication of work.

[cloud-j-luna]

The client libraries remove dependencies on cobra and others. Provides simple interfaces that use the upstream types from the protobufs. Has helper functions to reduce the amount of code necessary to interact with the network.

[troian]

I see, the existing client does not have cobra dependencies and works with protobuf types from akash-api

[cloud-j-luna]

I can have a deeper look at that client implementation. Do you believe such an implementation was not necessary? From our side we needed a simpler interface to ingrate with Akash and we chose to follow the specification, which we think the comunity might take value from. 😁

[troian]

The community would have benefited from a single well-defined client API rather than multiple.

Yes, please, look at the existing client as it already implements a simple interface to access the Akash blockchain without any heavy dependencies.

[cloud-j-luna]

Was the client that you linked the single well-defined client API? If yes I think there is no further discussion to be had. From our side we use it and has helped our code be more maintainable abstracting the complexities associated with operating Akash. We don't feel it was unecessary, on the contrary thats why we implemented it. I'm interested in the thoughts of the remaining community. Thank you for the feedback! 😁